DESCRIPTION: (Applicant's Abstract) This proposal seeks support for a ten-year follow-up of the 5877 people who participated in Parts I and II of the National Comorbidity Survey (NCS; Kessler et al., 1994), a nationally representative survey of people in the age range 15-54 in the U.S. household population designed to estimate the prevalences and correlates of DSM-III-R mental and substance disorders. The focus of the follow-up survey will be on the relationship between temporally primary mental disorders and the subsequent onset and course of substance disorders. Specifically, we aim to use information about mental disorders collected in the baseline survey to predict the subsequent onset of substance use and the onset and course of substance disorders over the intervening decade. We also aim to use causal modeling to evaluate the effects of mediators and modifiers of these associations. There are two specific aims. The first is to evaluate the two overarching hypotheses that guide the proposed research: that temporally primary mental disorders assessed at baseline predict the subsequent first onset and progression of substance use and substance disorders over the intervening decade; and that co-occurring mental disorders among baseline respondents with substance disorders are significant predictors of continuation over the intervening decade. The second specific aim is to carry out more refined analyses to elucidate the mechanisms involved in the associations documented under the first aim. A large number of potentially important mediators and modifiers will be considered here, with a particular interest in modifiable risk factors that could be intervention targets associated with mental disorders. All of the work under this second aim will separately consider the subsample of respondents with primary conduct disorder and the subsample of respondents without conduct disorder who have primary anxiety-mood disorders. The processes involved in the onset and course of secondary substance disorders are anticipated to be considerably different in these two subsamples, with substance disorders strongly related to a broader pattern of unconventionality and risk taking in the subsample with primary conduct disorder and to self- medication in the subsample with primary anxiety-mood disorders.